Content area
Full Text
Two studies of children's conversational language abilities are reported. In the first, mean length of utterance (MLU) and lexical diversity (D) were examined in a group of typically developing Cantonese-speaking children in Hong Kong. Regression analyses indicated a significant linear relationship between MLU and age (R = .44) and a significant curvilinear relationship between D and age (R = .73) in children age 27-68 months. MLU and D were moderately correlated with each other (r = .23); however, the two measures showed no statistical relationship when the effect of age was partialled out. In a second study, the utterances of Chinese children with specific language impairment (SLI) were found to be significantly shorter and less lexically diverse than typically developing children matched for age but similar to children matched for comprehension level. Discriminant analyses revealed that the combination of age, MLU, and D could be used to accurately differentiate children with SLI from both age-matched and language-matched children. Based on these findings, we hypothesize that these measures can be used jointly as a marker of SLI in Cantonese-speaking children.
KEY WORDS: child language development, child language disorders, specific language impairment, lexical diversity, Cantonese, MLU
Examining various aspects of children's language production by computing quantitative measures from samples of their spontaneous speech has a long-standing tradition in English. Two aspects in particular, utterance length and vocabulary diversity, have received more attention in the literature on normal and atypical language development than any other. In this paper we focus on their application to Cantonese-speaking children and provide evidence that together they can be used to differentiate children who have language difficulties from those who do not.
Measures of utterance length, particularly mean length of utterance (MLU) in morphemes, gained popularity among researchers following Brown's (1973) assertion that children of the same MLU were more alike in their language development than children of the same age. One outcome of this as far as researchers were concerned was that MLU came to replace chronological age as a metric for mapping out sequences of language development (e.g., Miller, 1981; Morehead & Ingram, 1973). This approach led to the standard research paradigm used in the field of child language disorders today by which a group of children with language difficulties...